Methodology

Disc roll values (RV)

An S-rank disc has 4 substats; each substat has a maximum single-roll value. Roll value is the share of perfect rolls your disc achieved:

RV% = 100 × Σ(valueᵢ ÷ maxRollᵢ) ÷ 4

Default max single-roll values (S-rank) [VERIFY per patch]: ATK% 3%, Flat ATK 19, HP% 3%, DEF% 4.8%, Crit Rate 2.4%, Crit DMG 4.8%, PEN 9, PEN Ratio 2.4%, Anomaly Proficiency 9.

Crit value (CV)

CV = 2 × Crit Rate% + Crit DMG%

The standard community convention for comparing crit sticks. Double-weighting crit rate reflects its scarcity on discs.

Damage model

hit  = ATK × multiplier × (1 + DMG bonus) × (1 − RES)
crit = hit × (1 + Crit DMG)
expected = hit × (1 + CR × CD)

Penetration, anomaly and stun-window multipliers are deliberately out of scope v1 — the model documents what it includes so results stay auditable.

Pull / pity model

Shared HoYo-style model [VERIFY]: base 5★ rate 0.6%, soft pity ramp from ~pull 74 (+6%/pull, community estimate), hard pity 90, 50/50 with carried guarantee. The pull planner exposes every parameter so you can adjust to your game's current behavior.

The [VERIFY] policy

Every game value carries a verification flag and the dataset carries a review date (currently 2026-10-08). We label uncertainty instead of presenting guesses as fact. Patch-day updates: re-check sources → update src/data/site.ts → bump the date → note in the changelog.

AI-assist disclosure

Pages were drafted with AI assistance and reviewed by the site operator before publishing. All math is deterministic, auditable client-side code.